Proposed: Comprehensive scenario demonstrating a multitude of different CDS use cases.

This scenario and its constituent uses cases are being developed by the CDS Collaboratory, a "joint venture" between the OpenCDS, Socratic Grid that will soon include other open source HSPC source Logica participants. 

"A VA patient with essential hypertension is traveling in the Phoenix area and is involved in a severe car accident that leaves him severely injured and comatose. He is admitted to a local intensive care unit and is being treated, among other things, with drug XXX for Metoprolol for hypertension. His vital signs and arterial blood pressure are displayed on a bedside CR Monitor. A Clinical Decision Support System (CDSS) monitors the patient's care and advises the clinical staff as they treat the patient for his injuries. 

Use Case #1: Several days into the hospitalization, the system detects that the patient is experiencing a transient, but clinically significant (suboptimal) drop in systemic blood pressure. These dips, however, are still within the patient monitor's configured parameters and so no alarm is triggered. the dips are correlated with with each XXX administrationMetoprolol administration, Typically, these transient events would be managed by a dose and/or interval adjustment if there is no risk of infection.  However, the hospital is a NwHN participant and has access to the patient's VA medical history and genome profile.  It determines that the patient is genetically predisposed to unpredictable responses to his XXX medication Metoprolol medication.  After determining alternative therapies, the system checks a drug-allergy knowledge base, evaluates the possibilities in light of the patient's VA allergies, recommends a suitable substitute and generates the corresponding provider recommendation.

Use Case #2: When the provider fails to acknowledge the medication alert within a mandated period of time, the system escalates the message to the provider's cell phone. Upon receiving an appropriate HIPAA compliant text, the provider logins in, opens the alert, reviews several InfoButton provided references, and ultimately accepts the recommendation. The system then retrieves the corresponding orderable from an Order Catalog and the provider completes and signs the order.

Use Case #3: Several days later the patient begins to experience blood pressure dips that are not temporally related to medication administrations. The system determines that these events should be evaluated by a third party sepsis screening system to investigate a possible infectious etiology.  This system responds with an appropriate alert notifying the provider that a sepsis work up is recommended.

Use Case #4:  Upon recovering from his injuries, and as the patient is preparing to be discharged, the system prompts the provider to complete a suicide risk assessment as recommended for all former combat veterans. Upon completing the survey, the system scores the completed screening tool and determines that the patient is a high risk.  In response, the discharging physician makes a mental health follow-up appointment and personally calls the receiving physician to ensure appropriate transition of care."

Architecture

Patient physiologic waveforms (created by a patient simulator) are displayed by bedside medical devices. These devices output data to a C4MI service that normalizes the structure and semantics of the data before passing these "real-time" events to the rest of the architecture.

The patient monitor data is analyzed by an Event Processor (providing complex event processing of waveform data) that extracts a variety of features from the event stream and publishes those features (facts) to an HL7 compliant Event Publication and Subscription Service.

The CDS system is a subscriber to the Medical Device topic in the EPS topic tree. The system also subscribes to many other topics that publish clinical events being recorded in the EMR. One of these topics is Medication Administration.

The CDS system can make requests to non-self knowledge bases to augment its own analytic capabilities - our use case illustrates an HL7 DSS call to OpenCDS to evaluate whether the patient may have sepsis. These systems can either publish their return "advice" to an appropriate EPS topic or can directly communicate back to the original CDS system.

The CDS system analyzes the data available, including any derived facts from third-party services, and publishes its final "advice" to EPS.  An HL7 compliant Unified Communication Services is a subscriber to the EPS Advice topic. It is responsible for communicating advice to recipients and managing any required re-routing or escalation.

The EMR manages provider visualization of advice in part by using a Smart-On-FHIR CDS "Inbox".  The Inbox is a Smart On FHIR component that plugs in to a CareWeb bedside GUI.  When the provider fails to acknowledge the alert within a mandated period of time, the system escalates the message to the provider's cell phone.  Upon receiving an appropriate HIPAA compliant text, the provider logins in, opens the alert, reviews several InfoButton provided references, and ultimately accepts the recommendation.

The CDS system can process and display HL7 HeD compliant structured documents, order sets and simple rules.  In our use case, the provider is presented with a Suicide Risk assessment.  The  system can evaluate a completed form/survey using an HeD provided rule, and notifies the provider that the patient is at high risk.